Speech pause distribution as an early marker for Alzheimer’s disease

Abstract

Background: Pause duration analysis is a common feature in the study of discourse in Alzheimer's disease (AD) and may also be helpful for its early detection. However, studies involving patients at the preclinical stage of mild cognitive impairment (MCI) have yielded varying results.Objectives: To characterize the probability density distribution of speech pause duration in AD, two multi-domain amnestic MCI patients (with memory encoding deficits, a-mdMCI-E, and only with retrieval impairment, a-mdMCI-R) and healthy control groups (HC) and to check if there are significant differences between them. To discuss the potential of those findings in clinical practice.Method: 112 picture-based oral narratives were manually transcribed and annotated for the automatic extraction of pause durations and their subsequent logconversion. We consider different probability distributions to fit speech pause duration truncating shorter ranges taking into account latest statistical findings to avoid inherent methodological uncertainties present in them.Results: Lognormal distribution (LND) explains the distribution of pause duration in speech for all groups, and its fitted parameters (µ,σ ) followed a gradation from the group with shorter durations and a higher tendency to produce short pauses (HC) to the group with longer pause durations and a considerably higher tendency to produce long pauses with more variance (AD). Importantly, a-mdMCI-E produced significantly longer pauses with greater variability than their a-mdMCI-R counterparts (α = 0.05) across all groups of study. Conclusion:We characterize and report significant differences at group level in the speech pause distribution across all groups of study that could be used to design tools and experiments for early prediction of AD progression.

Clinical classification of memory and cognitive impairment with multimodal digital biomarkers

“…The number of long pauses has evidence of being negatively impacted by cognitive impairment in a number of different speech contexts: reading, sentence repetition, picture description, and spontaneous speech. 25,53,55,56 The current research demonstrates increased number and length of silent pauses in a relatively short three-word recall context for individuals with cognitive impairment.…”

A Semi-Supervised Speech Deception Detection Algorithm Combining Acoustic Statistical Features and Time-Frequency Two-Dimensional Features

“…This process is accompanied by changes in ERPs on the amygdala, insula, and prefrontal regions of the brain as well as changes in acoustic signature parameters associated with lying, with some studies demonstrating that these two changes are correlated [23]. Drawing on the work of Low et al [47] and Pastoriza-Domínguez et al [48] who used machine learning algorithms based on acoustic feature analysis for detecting major mental disorders, we focused, in this paper, on choosing the acoustic feature parameters associated with the act of lying and used the trained neural network model to detect subtle changes in the acoustic feature parameters under different speech patterns to discriminate between lies and truth. This can help us better understand how speech is processed in the brain and enable researchers to further investigate the brain's cognitive neural mechanisms during the lying process.…”

Screening of Mild Cognitive Impairment Through Conversations With Humanoid Robots: Exploratory Pilot Study

“…We extracted 7 types of features concerning the speech intervals, such as the duration of the speech period and the proportion of fillers, using the speech interval and filler labels. The duration of the speech period, the response time, and the proportion of the silence time are features that previous studies have shown to be effective [ 12 , 15 , 17 , 18 ]. It is difficult to extract features such as the speech period and response time with sound-collecting microphones only when there are speech superimpositions.…”